Post

Transformers 4.54.0 is out! This release adds support for compute kernels hosted on the Hub. When enabled, transformers can replace PyTorch layer implementations by fast, specialized kernels from the hub. github.com/huggingface/...

github.comRelease v4.54.0: Kernels, Transformers Serve, Ernie, Voxtral, LFM2, DeepSeek v2, ModernBERT Decoder... · huggingface/transformersImportant news! In order to become the source of truth, we recognize that we need to address two common and long-heard critiques about transformers: transformers is bloated transformers is slow O...

26/7/25, 11:34 am2 reposts7 likes

Daniël de Kok

@danieldk.eu

26 July 2025

The kernel ecosystem is completely open: you can make your own kernels with kernel-builder, upload them to the hub, and register a mapping using the kernels package and they get used by transformers. github.com/huggingface/... github.com/huggingface/...

David Holz made an introduction video showing how to make your own kernels with kernel-builder: www.youtube.com/watch?v=HS5P...

youtube.comHugging Face Kernel Builder Walkthrough | Image to Grayscale CUDA KernelYouTube video by David Holtz